High Network Jabber, is it normal?

  • 1
  • Question
  • Updated 3 years ago
  • (Edited)
Having only just recently upgraded to some proper switches across our network we are just starting to get some real data.  One of the strange things I'm seeing is a rather large number of errors.  From a network monitor occasionally reporting more than 800 errors/sec and the switch itself reporting more than 15515965 jabbers in less than a month (This was just on one port, btw.  The total number is much higher).

The real question here is, is this normal?  It seems disturbingly high to me, but maybe this is normal behavior for an Aerohive AP.  Out of 96 ports, the only ones that seem to report errors are those connected to APs (Three of them all report varying degrees of errors) and the uplink port (Makes sense if all the APs are sending out erred packets).

We have a good deal more APs, but this was the first set of switches to get replaced and therefore to give us useful data.
Photo of Brandyn Baryski

Brandyn Baryski

  • 8 Posts
  • 0 Reply Likes
  • Confused

Posted 3 years ago

  • 1
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2491 Posts
  • 451 Reply Likes
Hi Brandyn,

Colour me skeptical, but if this is happening with all your APs, I would be more inclined to suspect a fault on the switch, perhaps just faulty accounting in the switch software.

Can you reproduce this with another vendor's managed switches?

What model(s) of APs do you have? What's the switch model?

And just out of curiosity... What cable grade are you using?

Is the link speed 1 Gb/s (1000BASE-T)?

Nick
(Edited)
Photo of Brandyn Baryski

Brandyn Baryski

  • 8 Posts
  • 0 Reply Likes
I don't really guarantee this is happening with all APs, just all on the newly replaced switch.

It is a Dell N2048P switch servicing three AP121.  The rest of our network is being being upgraded from netgear switches (which don't seem to report packet error in any meaningful way) to similar models over the next two to three months.  This was simply the first of the switches to go.

Note: I am getting similar readings from the switch software and from SpiceWorks Network Monitor.  The latter provided the 800 errors/sec and the former provided the number of jabber packets since it was installed.

Cabling is Cat 5e and the switch is auto negotiating and all three ports read at 1Gb/s.  

I can't really see anything that really seems out of the ordinary.  Unless packet size is an issue.  Do Aerohive APs normally send out packets larger than 1518?  I'd be perfectly inclined to believe this is just the switch complaining about over sized packets, modify the AP settings and call it a day.  But, I just find it strange that this is default behavior as I've never touched anything relating to packet size on configurations.
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2491 Posts
  • 451 Reply Likes
The monitoring will be via SNMP and reading, ultimately, the same counters.
Is the software on the switch up-to-date?
Have you tried enabling jumbo frames support to see if it makes a difference to what's reported?
(Edited)
Photo of Brandyn Baryski

Brandyn Baryski

  • 8 Posts
  • 0 Reply Likes
The switch firmware is the most recent version.  I can certainly enable jumbo frames and see if that does anything.  I was more or less interested in finding out if this was expected behavior.  I've never seen anything mentioning that jumbo frame support is needed.

In either case, I will enable the jumbo frames on the ports and see if that solves the issues.  As soon as I can figure out how to enable jumbo frames.

Edit: I ended up having to set the entire switch to have a larger MTU.  Not a solution I am expressly happy with, but this is merely to test to see if that helps to alleviate the issue.
(Edited)